Two Prevalent Listeria ivanovii subsp. ivanovii Clonal Strains With Different Virulence Exist in Wild Rodents and Pikas of China

Listeria ivanovii subsp. ivanovii is an intracellular bacterium distributed widely in nature, causing the listeriosis in ruminants and humans. Previous researches had isolated 116 strains of L. ivanovii subsp. ivanovii from wild rodents and pikas of different regions in China, and the predominant sequence types were ST1 and ST2. In this study, we first investigated the biological characteristics and virulence of these two clonal strains including motility, metabolism and virulence in cells and mouse model. The results demonstrated the ST1 strains exhibited motility, wide metabolic activity and hypervirulence, whereas the ST2 strains showed non-motility, relative lower metabolic activity and virulence. Considering the transmissible ability from wild rodents and pikas to ecological environment, the L. ivanovii subsp. ivanovii with potential pathogenicity to humans and ruminants should be monitored.


INTRODUCTION
Listeria ivanovii and Listeria monocytogenes, which are the only two pathogenic in the genus Listeria, appear to be widely distributed in a variety of environments such as food, water, soil and feces of humans and animals (1)(2)(3). L. monocytogenes can cause various clinical manifestations such as septicemia, meningoencephalitis and diarrhea in human and animals. L. ivanovii can cause ruminant infection with similar pathological manifestations except meningoencephalitis, which is one of the characters of L. monocytogenes infection in ruminants and humans (2,4,5). In recent years, sporadic listeriosis caused by L. ivanovii in immunocompromised people have been reported (5,6).
The two subspecies of L. ivanovii, which are subsp. ivanovii and subsp. londoniensis, had been reported (7,8). They shared the same biochemical characteristics except the ability to utilize N-acetyl-β-D-mannosamine and ribose, which can be used to identify the subspecies (8). The strains of L. ivanovii subsp. ivanovii are sensitive to bacteriophages infection. While the strains of subsp. londoniensis seem to be resistant to bacteriophage infection except for a large virulent A511-like virus, and the LivCRISPR-1 which is a type II-A system may play a native role in phage defense (9,10). L. ivanovii subsp. ivanovii caused the most of listeriosis in human and animals, and a previous study had found L. ivanovii subsp. ivanovii clinical strain G770 contains a type I restriction-modification system which may contribute to its hypervirulence (1,6).
A total of 116 L. ivanovii subsp. ivanovii strains were isolated from intestinal contents of wild rodents from 12 provinces in China during 2014-2016 (12,13). The strains belong to five sequence types (STs), and the predominant subtypes were ST1 (10/116, 8.62%) and ST2 (100/116, 86.21%). ST1 strains were isolated from feces of Mus pahari, Apodemus chevrieri, Apodemus draco and Niviventer confucianus in Tibet and Yunnan province, and their wide-range distribution suggests possible presence of bacteria transmission. ST2 strains were collected from intestinal contents of Ochotona curzoniae (pikas) and Marmota himalayana in Qinghai province. As the carriers of pathogens, wild rodents such as Rattus rattus and Rattus norvegicus have high reproducibility and free-living behavior, which enable them to play roles in ecological environment security by fecal-oral route (14,15). Wild rodents and pikas that can carry pathogens such as H5N1 highly pathogenic avian influenza virus and shiga toxin-producing Escherichia coli, may play role in pathogenic Listeria transmission among various species and even influence on ecological environment (16,17). However, no data was reported about the biological characteristics and virulence of different ST strains of L. ivanovii. In this study, we first investigated the motility, metabolism and virulence of ST1 and ST2 strains from wild rodents of China.

Average Nucleotide Identity (ANI) Analysis
The 28 strains (Table S1) of Listeria with genome sequences available online were chosen for ANI analysis (18). The unweighted pair group method with arithmetic mean method (UPGMA) based dendrogram was constructed by Software Bionumerics.

Motility Assay
The L. ivanovii subsp. ivanovii strains were inoculated in BHI semisolid medium at 25 • C and scanned by Transmission electron microscope.

Relative Expression Quantity of Genes
Real time-PCR was used to determine the relative expression quantity of LIPI-1 genes and motility related genes (gmaR, flaA). RNA was extracted from log-phase bacteria by Trizol (Ambion), then reverse transcribed by PrimeScript TM reagent kit with gDNA Eraser (TaKaRa) according to manufacturer's instructions. The acquired cDNA was measured by Real-time PCR using SYBR R Premix Ex Taq TM II (TaKaRa) and calculated by 2 T − C method (19). Each RT-PCR was conducted in the 20 µL aliquot of running buffer containing 50-100 ng cDNA, annealing temperature was 60 • C. The relative expression quantity was normalized by reference gene 16SrDNA. Primers were shown in Table S2.

Recombination Plasmids Construction and Electro-Transformation
To construct recombination plasmids, the DNA fragments were obtained by PCR using primers in Table S2 and purified DNA Gel Extraction Kit (Takara). The purified products were cloned into restriction enzyme cutting sites of vector PIMK2 by T4 ligase (NEB) cloning. Recombination plasmids were introduced into competent cells of LIV047, and the electroporation conditions were 2.5 KV, 25 µF, 400 . Then the recombination strains were screened from BHI plates containing 50 µg/mL kanamycin.

In vitro Growth Assay and Phenotype MicroArray (PM) Analysis
For in vitro growth assay, the log-phase bacteria were transferred to BHI broth with 1:100 dilution and cultured for 72 h, and the OD600 values were determined by Bioscreen C microbiology reader per hour.
The Carbon source (C-source, PM01-02) and Nitrogen source (N-source, PM03, 06-08) utilization tests were performed as described in Biolog instructions of Phenotype MicroArray Procedures for B. subtilis and other GP Bacteria (20). The areas of growth curves (72 h) of L. ivanovii subsp. ivanovii strains were compared.

Identification of Virulence Related Metabolism Genes of Differential Metabolism Phenotypes Between ST1 and ST2 Strains
The metabolism genes related with virulence were systemic summarized by previous studies (21)(22)(23)(24). The similarities of amino acid between the tested strain and PAM55 were calculated by software MegAlign.

Invasion and Intracellular Growth Assays in Caco-2 and MDBK Cells
Human-derived Caco-2 (2 × 10 5 ) and bovine-derived MDBK cells (1 × 10 5 ) were incubated in 24-well plates the day before invasion with the log-phase bacteria at multiplicity of infection (MOI) 100. After an hour infection, cells were incubated for 1 h in DMEM (for Caco-2) or MEM (for MDBK), both containing 50 µg/mL gentamicin. Then the bacteria were released from cells through adding PBS containing 0.1% Triton-X100 at 1, 3, 6, and 12 h post-infection, and incubated on BHI plate for colony counting.

Cytotoxicity Assays in Caco-2 and MDBK Cells
Cytotoxicity assay (Promega) was used to measure cells lactate dehydrogenase (LDH) release that indicates the cells membrane damage. Caco-2 (4 × 10 4 ) and MDBK cells (2 × 10 4 ) were incubated in 96-well plates the day before infection with the bacteria at MOI 100. After 3 h of incubation, the LDH in cell supernatant was measured according to manufacturer's instructions (25).

Mouse Infection and Examination of Organs
Experimental procedures involving mouse model were approved by Welfare and Ethical Review in Animal Experimentation of National Institute for Communicable Disease Control and Prevention, China CDC, and carried out by the licensed individual with a license number of 2017-0021. Six to seven weeks old female BALB/c mice (Charles River) were used in the assays and conducted as previous described with some modifications (26)(27)(28). For 50% lethal doses (LD 50 ) determinations, groups of 10 mice were injected intraperitoneally with 5-fold dilutions of log-phase bacteria (5 × 10 8 , 1 × 10 8 , 2 × 10 7 , 4 × 10 6 , and 8 × 10 5 CFU in 200 µL PBS) and monitored for 7 days, mice of control group were injected with 200 µL PBS.
The infection doses were 10-fold lower than LD 50 (1.70 × 10 6 CFU) of the reference strain PAM55 in subsequent experiment analyses. Five mice were euthanized for weighing and analysis of bacterial loads in organs (liver, spleen, lung, kidney, small intestine, and brain) each for seven consecutive days, and histopathological changes of organs were analyzed at 1, 4, and 7 days post-infection (dpi). Each organ was divided into two parts, one for hematoxylin-eosin (HE) staining, the rest was homogenized in PBS containing 0.5% Triton-X100 and 10-fold diluted for colony counting with three replicates on BHI plate. In addition, small intestine tissues were rinsed by DMEM and incubated in DMEM containing 100 mg/L gentamicin for 1 h to kill extracellular bacteria before HE staining and colony counting.

Identification of Virulence Genes
The genome sequence of ST1 strain and ST2 strain were analyzed on website of Virulence factors of Pathogenic Bacteria (http:// www.mgc.ac.cn/VFs/main.htm), and the similarities of amino acid were calculated by software MegAlign (29). PAM55 was used as reference strain.

Statistical Analyses
Data of curves were presented as mean ± SD and subjected to repetitive measurement deviation analysis. Other results were presented as mean ± SD and subjected to Two-way ANOVA and Student's test. Differences were judged statistically (SPSS 20.0) significant at P < 0.01 or 0.05. Software Graphpad 7.0 was used to plot figures.

RESULTS
The 20 species of Listeria including L. ivanovii were used to construct a UPGMA based dendrogram (Figure 1). The reference strain PAM55, ST1 strain LIV037, and ST2 strain LIV047 formed an independent clade and the ANI between them was larger than 99% (30).

Motility Analysis
The reference strain PAM55 and ST1 strains demonstrated cloudy growth and swarming motility around the stab in semisolid medium at 25 • C. No evidence of motility was observed in all ST2 strains (Figures 2A-D). The flagellum was observed in ST1 strain LIV037 using transmission electron microscope, but not in ST2 strain LIV047. Compared with strain PAM55, 24 motility related genes in LIV047 strain have non-synonymous mutations and/or indels (Table S3). Notably, in ST2 strains, a premature stop codon (PMSC) was identified in gmaR, which was important in regulating expression of flagellar genes such as flaA, and the relative expression levels of gmaR and flaA in LIV047 were significantly lower than that in PAM55 and LIV037 (P < 0.01) using real-time PCR (Figure 2E). However, a recombinant strain LIV047-cgmaR with full length gmaR in constitutive overexpression vector still showed non-motility (data not shown), and the expression levels of flaA was still significantly (P < 0.01) lower than ST1 strains.

Metabolism Phenotypes
The proliferation ability of strains was compared by growth curve in BHI broth. Significantly different growth rate (P < 0.01) between ST1 and ST2 strains were shown in Figure 3, and exponential phase of ST2 strains lagged behind ST1 strains by about 3 h, and ST1 strains shared similar growth characteristics with PAM55 ( Figure 3A). The utilization of Csource and N-source were tested by Phenotype MicroArray system (Figures 3B,C). The comparison of growth curves in 192 FIGURE 1 | The dendrogram of UPGMA clustering based on ANI. The 20 species of Listeria including L. ivanovii were used to construct the dendrogram. The vertical bar represents the 95% ANIb that correlates with the 70% DNA-DNA hybridization threshold (30). The reference strain PAM55, ST1 strain LIV037, and ST2 strain LIV047 were formed an independent clade respectively, and the ANI between them were larger than 99%.
C-sources (PM01-02) showed that ST2 strain LIV047 had a lower metabolic activity. No significant difference was found between ST1 strain PAM55 and LIV037 regarding C-sources utilization. In the 384 N-sources (PM03, 06-08) utilization assays, ST1 strain LIV037 grew better than other strains, and ST2 strain LIV047 also exhibited a lower metabolic activity.
In addition, 11 carbon and nitrogen differential utilization related proteins were analyzed, and the similarities of proteins in LIV037 and LIV047 refer to PAM55 were displayed in Table S4. Most proteins in LIV037 and LIV047 have bits of non-synonymous mutations (similarity >98%). However, glycerol utilization related protein glpF had 105 non-synonymous mutations, and ethanolamine utilization related gene eutA, eutB, and eutC were absent in the ST2 strain LIV047, which was consistent with the result of glycerol and ethanolamine utilization in Phenotype MicroArray test.

Virulence Analysis
The ability of internalization, bacterial cytotoxicity, and intracellular growth were evaluated in both Caco-2 and MDBK host cells. Significantly (P < 0.01) higher invasion ability and cytotoxicity were found for ST1 strains than ST2 strains in both Caco-2 and MDBK cells (Figures 4A,B). For intracellular growth assay, ST1 strains exhibited significantly (P < 0.01) more efficient growth rate than ST2 strains, and the rapid growth appeared obviously from 3 to 6 h post-infection in Caco-2 and MDBK cells (Figures 4C,D).
The virulence of ST1 and ST2 strains was examined in vivo using BALB/c mouse model of intraperitoneal infection. The LD 50 of PAM55, LIV037, and LIV047 to BALB/c mice were 1.70 × 10 7 CFU, 2.00 × 10 7 CFU, and 2.24 × 10 8 CFU, respectively. Then the virulence of strains was further evaluated in mice infected by ST1 and ST2 strains (1.70 × 10 6 CFU), changes Frontiers in Veterinary Science | www.frontiersin.org  in body weight and bacterial loads in organs were measured for 7 days post-infection. No significant difference (P > 0.05) of maximum weight loss was observed between mice infected with PAM55 (1 dpi, 3.74% of original weight) and LIV037 (2 dpi, 4.38% of original weight); however, significantly less weight loss (1 dpi, 1.90% of original weight) was observed in mice infected with LIV047 than that of PAM55 (P < 0.05) and LIV037 (P < 0.01) (Figure 5A). Bacterial loads in liver, spleen, lung, kidney, intestine, and brain of mice were measured daily. No bacterium was detected by colony counting and real-time PCR (genus-specific gene prs) in the brain tissues of all mice (data not shown), the bacterial loads in other organs of mice of LIV047 infected group were significantly (P < 0.01) less than that of PAM55 and LIV037 infected groups. The bacteria were persistent in the liver, spleen, lung, kidney, and intestine of the L. ivanovii infected mice through 7 days except for the LIV047 group in which bacteria disappear at 5 dpi in liver, 4 dpi in lung and 6 dpi in kidney (Figures 5B-F).
To evaluate and compare pathological injury in mice organs, mice were sacrificed and examined at 1, 4, and 7 dpi. Consistent with the bacterial loads, PAM55 and LIV037 strains showed more severe pathological injuries to mice than LIV047. In the mice with PAM55 and LIV037 infection, liver cell degeneration and necrosis with lots of neutrophils aggregation emerged in livers at 1 dpi, and severe necrosis appeared at 4 dpi, which lasted until 7 dpi for PAM55 group, while only a few of neutrophils were observed for LIV037 group. In contrast, the necrosis only appeared in liver tissue at 1 dpi for LIV047 group, and few inflammatory cells existed at 4 and 7 dpi (Figures 5G-J,  Figure S1). Marked neutrophil infiltration were observed in red pulp of the spleen at 1, 4, and 7 dpi for mice of PAM55 and LIV037 infected groups, but 1 dpi for LIV047 group with fewer neutrophils at 4 and 7 dpi (Figures 5K-N, Figure S2). Wild inflammation including hemorrhage, alveolar septum widened, collapsed alveoli and inflammatory cells infiltrate were observed in the lungs of mice infected by PAM55, LIV037, and LIV047 at 1 and 4 dpi; same character also lasted at 7 dpi for LIV037 and PAM55, but not for LIV047 (Figures 5O-R, Figure S3). No histopathology change was observed in kidney of all three groups (Figures 5S-V, Figure S4). For small intestine, the granuloma was observed in the mice infected by PAM55 and LIV037 strain at 4 and 7 dpi, and by LIV047 strain at 7 dpi (Figures 5W-Z, Figure S5).
The virulence genes of strains were analyzed. ST1 strain LIV037 and ST2 strain LIV047 were positive for LIPI-1, LIPI-2 and other 26 virulence genes, and LIV037 has more similarity to PAM55 than LIV047 (Table S5). Compared to PAM55, most virulence factors in ST1 and ST2 strains have bits of nonsynonymous mutations (similarity >95%). However, in LIV047, deletions and several non-synonymous mutations were identified in actA (119 amino acids deletion which located on an unknown domain and 15 non-synonymous mutations), inlB2 (309 amino acids deletion which located on the bacterial adhesion/invasion protein N-terminal and Gly-Tryp dipeptide domain, and 3 nonsynonymous mutations) and agrC (237 amino acids deletion which located on the hyaluronan mediated motility receptor Nterminal, sensor kinase SpoOB-type and alpha-helical domain). In addition, LIV037 and LIV047 shared similar expression levels of six virulence genes in LIPI-1 with PAM55 strain (Figure 6A), but expression levels of plcA gene coding 1-phosphatidylinositol phosphodiesterase in LIV047 was significantly lower than that of PAM55 and LIV037 strains (P < 0.01). Four amino acid changes (S83A, K235E, A245T, and T310R) in plcA were identified for LIV047 strain compared with PAM55 strain. Then a recombinant strain LIV047-cplcA containing the same plcA in PAM55 strain was constructed, but no significant difference of cytotoxicity was found between LIV047-cplcA and LIV047 strains (P > 0.05, data not shown). However, white halos around the colonies on the BHI plate containing 5% lecithin was observed for the recombinant strain LIV047-cplcA but not for ST2 strains including LIV047 (Figures 6B-E).

DISCUSSION
Listeria ivanovii was distributed widely in nature and has potential to cause listeriosis in ruminants and even immunocompromised humans (5,6). Previous studies have noted that wild rodents and pikas in China carried various STs of L. ivanovii subsp. ivanovii, and the predominant sequence types were ST1 and ST2 (12,13). ST1 strains were collected from five species of wild rodents in Tibet and Yunnan province, which reveals the potential ability of ST1 clonal strains to spread cross species and regions (13). However, the ST2 strains were only found in Yushu Tibetan autonomous prefecture of Qinghai province. This is the first study to characterize two different clonal STs of L. ivanovii subsp. ivanovii on the aspects of motility, metabolism and virulence.
Flagellar motility enhances extracellular survival and colonization in host during Listeria infection (31)(32)(33)(34). L. ivanovii subsp. ivanovii ST2 strains were the first identified non-motility subtype in this study. Previous report noted gene gmaR is the key thermometer which controls the temperature-dependent transcription of motility genes such as flaA in L. monocytogenes, and a premature stop codons (PMSC) contributed to the lack of Tetratricopeptide Repeat and Anti-Repressor domains was found in gmaR of ST2 strains (35). However, constitutive expression of complete gmaR did not recover the motility and expression level of flaA in ST2 strains. The changes of other motility related genes or other unknown mechanism involved in motility may probably attribute to the non-motility of ST2 strains.
The virulence analysis strongly supports the more virulent property of ST1 strains compared to ST2 strains. Transcription levels of six genes (actA, hly, mpl, plcA, plcB, and prfA) in LIPI-1 were measured for the two different clonal strains, of which plcA in ST2 strain was significantly lower than that of ST1 strain. Gene plcA was prfA-regulated, and cooperate with plcB to lyse the phagocytic vacuole membrane (36). By utilizing phosphatidylcholine, plcA contributes to the formation of white halo around colony in lecithin plate, and has been used as a virulence marker to distinguish the pathogenic and nonpathogenic Listeria to humans (37,38). The low transcription level of plcA in ST2 strain corresponds to lose of white halo. In addition, previous carbon-source regulation research found that cellobiose could repress the transcription of plcA independent of prfA pathway (39). Cellobiose could not be utilized by ST2 strain in metabolism assay, its accumulation in organism probably affect the expression level of plcA. We found complete plcA enabled LIV047 strain to acquire phenotype of white halo on plate, but no significant difference of cytotoxicity was observed between LIV047-cplcA and LIV047 in cultured cells. Other virulence genes analysis found deletion mutations and lots of non-synonymous mutations which could potentially affect genes functions existed in actA, inlB2, and agrC of ST2 strain LIV047, these virulence genes are important for pathogenicity of Listeria (40)(41)(42). Whether it is the reason for the lower virulence of ST2 strains needs to be proved.
Listeria has its specific metabolic adaptations to promote colonization in eukaryotic cells, and metabolic pathways of glycerol, glutamine, amino acids, and peptides are crucial for the evasion of innate immunity, intracellular replication and survival (22,24). We found ST1 strains exhibited higher carbon and nitrogen metabolic activity and growth rate in nutrient rich medium than ST2 strains, which was possibly associated with the wide distribution of ST1 strains in China.
The capacity of pathogens to efficiently utilize various nutrient sources in nature is crucial for its increased survival and expanded niche range in hosts and ecological environments (20). Glycerol utilization related genes such as glpF and glpK were upregulated in host cells and the intestine of mice (43,44). However, the glpF of ST2 strain has only 54.94% similarity to ST1 strain PAM55. Ethanolamine and branched chain amino acids (BCAAs) were required for intracellular replication, promote Listeria lipid membrane homeostasis and resistance to intracellular stresses (45,46). We found that eutABC genes required to ethanolamine utilization were absent in ST2 strain. Lower metabolic activity of utilizing glycerol, ethanolamine, amino acids, and peptides in ST2 strain are consistent with the lower virulence in cells and mouse models in this study. Whether the mutations or transcriptional levels of these carbon and nitrogen utilization related genes affect the virulence ST2 strain are unknown.
All ST2 strains were isolated from Qinghai-Tibetan, which is the highest altitude in the world and comprises unique ecological environment with intense ultraviolet radiation and poor carbon and nitrogen sources. Whether the extreme ecological environment contributed to the variation in genes especially for those metabolism related genes needs to be further explored. Besides, ST2 strain's carriage is high in wild rodents and pikas, possibly because the relative lower virulence of ST2 strains enables them to colonize the gut for a long period of time, further studies are needed to verify this hypothesis.
In conclusion, we first reported the phenotypic characteristics and virulence of two prevalent L. ivanovii subsp. ivanovii clonal strains in wild rodents and pikas of China. They varied from motility, metabolism to virulence in cells and mouse model. The ST1 strains exhibited motility, wide metabolic activity and hypervirulence, whereas the ST2 strains showed non-motility, relative lower metabolic activity and virulence. Previous study has noted that bovine trophoblastic cells are susceptible to both L. monocytogenes and L. ivanovii infection, which may explain the abortions and reproductive failures caused by L. ivanovii infection in cattle (47). Considering the transmissible ability from rodents to ecological environment, the L. ivanovii subsp. ivanovii strains especially for ST1 strains with potential pathogenicity to humans and ruminants should be monitored. Further studies should focus on the specific mechanism of adaption to the different habitats and virulence of the two clonal strains of L. ivanovii subsp. ivanovii.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention.

AUTHOR CONTRIBUTIONS
LG and CY designed the research study and wrote the manuscript.
LG analyzed the data.